Moskovitz, T;
Arbel, M;
Huszar, F;
Gretton, A;
(2021)
EFFICIENT WASSERSTEIN NATURAL GRADIENTS FOR REINFORCEMENT LEARNING.
In:
Proceedings of the 9th International Conference on Learning Representations: ICLR 2021.
ICLR: Virtual conference.
Preview |
Text
1830_efficient_wasserstein_natural_.pdf - Accepted Version Download (1MB) | Preview |
Abstract
A novel optimization approach is proposed for application to policy gradient methods and evolution strategies for reinforcement learning (RL). The procedure uses a computationally efficient Wasserstein natural gradient (WNG) descent that takes advantage of the geometry induced by a Wasserstein penalty to speed optimization. This method follows the recent theme in RL of including a divergence penalty in the objective to establish a trust region. Experiments on challenging tasks demonstrate improvements in both computational cost and performance over advanced baselines.
| Type: | Proceedings paper |
|---|---|
| Title: | EFFICIENT WASSERSTEIN NATURAL GRADIENTS FOR REINFORCEMENT LEARNING |
| Open access status: | An open access version is available from UCL Discovery |
| Publisher version: | https://openreview.net/forum?id=OHgnfSrn2jv |
| Language: | English |
| Additional information: | This version is the version of record. For information on re-use, please refer to the publisher’s terms and conditions. |
| Keywords: | reinforcement learning, optimization |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Life Sciences UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Life Sciences > Gatsby Computational Neurosci Unit |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10167378 |
Archive Staff Only
 |
View Item |
